What problem does it solve? AI agents working with Unity lose track of what changed while they were away and cannot safely drive the Editor without parsing fragile scene YAML. This Skill exposes structured operations to observe and control the Unity Editor so agents can resume work with accurate context. ## Core Features & Use Cases - Change Tracking: Read a persistent journal of scene, file, undo, and lifecycle changes via editor_get_changes with cursor-based paging, instead of parsing .unity YAML. - Play Mode Control: Enter, pause, step frame-by-frame, capture runtime errors and screenshots, and exit play mode with async job monitoring. - Editor State & Selection: Inspect live GameObject runtime state, get/set selection, read editor context, undo/redo, and execute exact menu items like File/Save. - Use Case: After the user edited the scene while the AI was offline, call editor_get_changes to rebuild context, then enter play mode with editor_play_capture to verify the game runs without errors. ## Quick Start Ask the agent to check the current Unity editor state and list any scene changes since the last session.